CEPA eprint 1389 (EVG-099)
Steps in the construction of “others” and “reality”
Glasersfeld E. von (1986) Steps in the construction of “others” and “reality”. In: Trappl R. (ed.) Power, autonomy, utopias: New approaches toward complex systems. Plenum Press, London: 107–116. Available at http://cepa.info/1389
Table of Contents
The notion of knowledge
“fit” instead of “match”
The initial construction of reality
Understanding of conceptual fit
Others and the notion of objectivity
A constructivist approach to ethics
The author advocates a change of perspective concerning the concept of knowledge. He suggests that the experiential reality in which we live and in which our sciences operate is the result of a self-regulating organism’s construction and should not be confounded with the ontological reality that philosophers have vainly searched for throughout the history of Western epistemology. The notion of viability gives a new slant to an instrumentalist theory of knowledge that serves as a basis for the cognitive construction of Others and may ultimately provide a starting-point for the development of a constructivist ethics.
Much of what I am going to suggest in this essay goes against notions that are widely held but rarely examined for their hidden presuppositions. We were all taught that there is virtue in “objectivity” and that it is sinful to question whether objective knowledge can ever be attained. Socrates, who proclaimed that it could not, was promptly put to death. The skeptics, who ever since then have maintained the impossibility of “true objective knowledge”, did not make themselves popular either. The philosophical profession, having been unable to counter their arguments satisfactorily, nowadays tends to consider them a persistent nuisance.
In this paper I shall present a few key characteristics of a theory of knowledge that is subversive in that it takes the skeptics seriously, but is also constructive in that it provides a working hypothesis that permits a non-contradictory analysis of the activity of knowing. Above all, I shall try to show that the experiential reality in which we live and in which our sciences operate should be considered the result of self-regulating construction and should not be confounded with the ontological reality which most philosophers, be they realists or idealists, bourgeois or marxist, have been searching for.
The notion of knowledge
In the philosophical tradition of the Western world, the concept of “knowledge” has almost without exception been understood to imply that the structures that result from cognition must in some way correspond to an external reality: “true” knowledge was supposed to depict or replicate what is real; and “reality” was intended to refer to a world “in-and-for-itself”, a world that exists ready-made, fully-structured, and independent of any cognizing subject.
At the beginning of that tradition, the skeptics had already pointed out that this conception of “knowledge” leads to a paradox. In order to test the required match between such knowledge and what it was supposed to be knowledge of, the experiencer would need some other access to the postulated reality; and that access would have to be immediate, so as to bypass the subject’s activity of knowing. Within the realm of the rational, however, no such immediate access seemed to be logically possible. This impasse has not been resolved in the course of philosophical history since then. In spite of countless attempts the paradox is as solid as ever, because the established conception of what “knowledge” ought to be has remained the same throughout. There were, of course, individual dissenters, such as Montaigne, Mersenne, Vico, and a few others, who realized that knowledge did not and could not live up to the general wishful expectation. Although these thinkers made valiant efforts to revolutionize epistemology, they had little impact on the tradition.
Only with the advent of control theory did it become possible to conceive of models of organization and government inside an organism and to view the cognitive enterprise as an outcome of self-regulation. Insofar as these models are self-contained with regard to cognition or “information”, they open a new perspective on epistemology. This possibility was slow to be realized because control theory was for the most part developed by engineers for whom the “feedback loop” was simply a powerful tool to construct highly efficient gadgets to which one could delegate certain tasks of guidance or control (Powers, 1978). These gadgets manifested goal-directed action, but the goals they pursued were of course the goals of the engineers who designed them (see Pask’s distinction of purpose of and purpose for, 1969).
The actual mechanical successes during this early infancy of cybernetics did much to obscure the possibility of applying the new concepts of circular causality and equilibrium in self-contained systems to the modeling of living organisms, the phenomenon of evolutionary adaptation, and, ultimately, to the problems of cognition.
It is well to remember that Jean Piaget, the most epistemologically oriented of the modern psychologists of cognition, formulated the core of a cybernetical theory of knowled3e when he wrote more than a decade before the official birth of cybernetics: “Intelligence organizes the world by organizing itself” (Piaget, 1937, p.311). This clearly and uncompromisingly created a new perspective on cognition and placed the emphasis on two hitherto neglected aspects: self-regulation and the endogenous construction of knowledge.
“fit” instead of “match”
Once the cognizing subject is no longer seen as a passive receiver of “information”, there is a radical shift of orientation. Perhaps the most dramatic consequence of that shift concerns the concept of “knowledge”. Instead of the paradoxical requirement that knowledge should reflect, depict, or somehow correspond to a world as it might be without the knower, knowledge can now be seen as fitting the constraints within which the organism’s living, operating, and thinking takes place. From that perspective, then, “good” knowledge is the repertoire of ways of acting and/or thinking that enable the cognizing subject to organize, to predict, and even to control the flow of experience. From this changed point of view, then, the cognitive activity does not strive to attain a veridical picture of an “objective” world (a goal which, as the skeptics have always told us, is unattainable), but it strives for viable solutions to whatever problems it happens to deal with.
This shift in the conception of knowledge is radical in more than one way. Not only is the notion of an absolute “truth”, a truth that matches ontological reality, abandoned, but with it the notion that each problem can ultimately have only one “true” solution must be given up. Unlike the conventional concept of “truth”, the concept of “viability” is not exclusive but reflects the common experience that the problems we face have, as a rule, more than one solution. Of course, this does not mean that all solutions to a problem must be considered equal. On the contrary, on a higher level of operating, where speed, economy, and even aesthetics are considered relevant factors, a solution may cease to be adequate, not because it does not attain the goal, but because it is too slow, too costly, or too cumbersome. The main conceptual shift, however, is in relinquishing the idea that true knowledge should be a veridical picture of an objective world.
Warren McCulloch said, in his 1948 lecture at the University of Virginia, that the break-down of a hypothesis is “the peak of knowledge” (McCulloch, 1965, p.154). It was a declaration of the break-down of traditional epistemology. When a plan of action or a conceptual structure (such as a hypothesis) fails, then and only then may we say that we have made contact with “reality” in the traditional sense. That contact, however, is at best a clash between our acting or thinking and the constraints within which our acting and thinking must take place. Such clashes may tell us something about our ways of acting and thinking, but they cannot provide a picture of the “real” world; they merely provide an indication of the insufficiency of the particular way of acting or thinking we have embarked on. When, on the other hand, an action or conceptual structure turns out to be successful, it tells us that we have remained within the constraints, have found a way that does not clash with anything and this, again, can provide no picture of the “real” world, because, in this case, we have done no more than act or think in the space the “real” world left unencumbered and free for us to act or think.
The fact that the concept of knowing was entangled with ontology and the notion of being from the very start of epistemological enquiry (cf. Plato’s Meno or Theaetetus) has had profound consequences. One of them was to make it exceedingly difficult to expound a theory of knowledge that cuts loose from “existence” and focuses exclusively on the cognitive activity and its results. The perennial idea that the knower, even if logic shows it to be impossible, must at least try to discover what the world is “really” like , is an idea that is not easy to discard. Although I believe that this idea must be discarded, I shall not press the point here. Instead I would ask you to consider this: if there is to be a knowing subject who can acquire “knowledge”, there must also be available some raw material, some “basic elements” out of which that knower can compose the structures which he or she is going to call “knowledge”. Such raw material or basic elements are usually supposed to be “data” or “information” that is conveyed to the knower from the “outside” via the senses. If, for the moment, we accept that realist hypothesis, it does not alter the fact that it is still the knowing subject who has the task of composing the data or interpreting the information in order to achieve a “representation” of reality. Whichever way one looks at it, therefore, “knowledge” cannot be a commodity that is found ready-made but must be the result of a cognizing subject’s construction. It is this constructive activity that we shall now look at a little more closely.
The initial construction of reality
What we ordinarily call “reality” is, of course, the reality of the phenomenal, the reality of the relatively durable perceptual and conceptual structures which we manage to establish, use, and maintain in the flow of our actual experience. This experiential reality, no matter what epistemology we want to adopt, does not come to us in one piece. We build it up bit by bit, and the construction is achieved by a succession of steps that come to form a succession of levels.
Repetition is an indispensable factor in that development. A simple sensory impression, a flash of color, for example, remains a dubious experience if we are unable to make it recur. Similarly, our concept of “existence” (in the sense of an experiential item “being there”, in its own right and independent of our experiencing it) is indissolubly tied to a notion of “permanence”; and permanence, after all, can be conceived only on the basis of at least two moments of experience, two moments that can be linked to constitute a continuity or to frame an interval during which the experienced item could be said to perdure. Such links have to be made and, in order to be known, they have to be made by the knower.
To a realist, it may sound absurd to say that repetition or recurrence has to be constructed by the experiencing subject. To recognize something as having occurred before is so commonplace, so “natural” an experience that it is easy to ignore (though the word clearly indicates it) that recognition is a cognitive activity and, as such, requires some doing. In this context it is essential to remember that I am speaking of epistemology, i.e. of knowing and not of being. In order to conclude that p, is the same as B, a comparison must be made and this comparison must yield the result of “sameness” rather than “difference”; and comparisons do not make themselves, they have to be made by an active agent.
The notion of “sameness”, without which we could not know that A is a repetition of B, is itself a much more complicated affair that it seems at first sight. Apart from the fact that it is always the cognizing subject who chooses the property or dimension in which two items are considered “the same”, the very concept of “sameness” involves an ambiguity that is of fundamental epistemological importance.
An example may help to make this ambiguity transparent. From the two statements “Julia bought the same dress as Ann” and “Julia slept in the same room as Ann” we would normally infer that, respectively, they imply two dresses but only one room. Of the two dresses we are told that they are equivalent in all respects one usually considers when comparing dresses; of the room, on the other hand, we assume that it is one and the same. The expression “the same”, thus, leads us to construct a relation of equivalence in the one case, and a relation of individual identity, in the other. Conceptually, the two relations are quite different: the one constitutes the basis for the formation of classes, i.e., collections of items that are considered “the same” in some respect; the other constitutes the basis for the construction of what Piaget has so aptly called “object permanence”, i.e. the notion that things have a life of their own and exist even during those intervals when they are not within the subject’s immediate experiential field. For a fuller exposition of this conceptual analysis, see Glasersfeld, 1984.
To continue with the “levels of reality”, a somewhat higher level is achieved whenever we are able to coordinate an experience in one sensory mode with an experience in another sensory mode. If an item (e.g. a patch of color) which we have isolated in our visual field can be recurrently “corroborated” by tactual exploration (e.g. a palpable edge), or when its appearance or disappearance can be coordinated with an auditory experience, then that item will be considered a good deal more “real” than if it remained exclusively visual.
There are many shades and subtle degrees in this construction of reality. The stars became more real once we were able to plot their motion, and the moon became more real to the man who stepped on it. But whatever reality each one of us creates for him- or herself is still precarious, because we all may have experiences that contradict our ordinary construction. We are subject to sensory illusions, have dreams, and sometimes experience hallucinations. At times, these irregularities cast doubts upon the reliability of our senses.
Fortunately we have another, much more powerful method of confirming the reality of our experiential world: the corroboration by Others. The way this method works, however, is, again, a great deal more complicated than it may seem at first sight. If I were to ask my neighbor for a carafe of water on the table in front of us, and he passed it to me, this would at once allay whatever doubts I might have had about the reality of the carafe. Things seen by several observers are taken to be more reliably “real” than things seen by only one. Very often, however, this corroboration by Others is taken to imply much more than simple experiential compatibility. Indeed, it is usually assumed that if you and I agree that we perceive, say, this carafe, our agreement could demonstrate not only that we have compatible experiences but also that what we experience must be a “true” reflection of what exists in an independent reality, whether we experience it or not.
In other words, the probability that Others experience something that seems compatible with what we ourselves experience, is usually taken as evidence that the “shared” experience reflects an independent “objective” reality. But David Hume, more than two centuries ago, showed quite conclusively that such inferences are based on faith, not on logical necessity.
Now, however, I would claim that these inferences or assumptions are illusory on two further counts, because both the process of communication and the Others with whom one communicates are not as straightforward as they may appear. In the two sections that shall follow I shall argue that neither communication nor the Others with whom we populate our experiential world can bring us closer to knowing an ontological reality.
Understanding of conceptual fit
Though corroboration by Others can be obtained without speaking (in the above example, I might simply have pointed at the carafe and the neighbor might have passed it wordlessly), in the overwhelming majority of cases corroboration is obtained by means of linguistic interaction. Language, thus, becomes an almost indispensable instrument in the construction of a “shared” reality. But here, once more, a confusion concerning the use of the expression “the same” creates the illusory impression that the shared reality “exists” and, therefore, must be independent of the communicating subjects who happen to share some knowledge of it.
To show that this is an illusory notion, we have to look more closely at how we come to “understand” language, communicatory gestures, or any other conventional semiotic system. Among other things, semiotic systems involve the formation of associations which link certain auditory or visual percepts (signals, signs, words, etc.) with specific other segments of our individual experience. The particular “meaning” of these communicative items is established gradually through interaction, in that the segments of experience a subject has associated with a particular sign or word are modified and adjusted until they fit into the situational contexts of recurrent communication events. When we say that we understand a piece of language, we are in fact saying that we are able to fit our interpretation of that piece of language into our assessment of the situation in which it was uttered, as well as into such as we might make of the speaker’s or writer’s intentions. All of this is and remains part of our own experiential world, the world that we have conceptualized during our past experience, and does not and cannot attain at any point a “real” world that is supposed to be independent of the language-user’s conceptualizations.
Language, thus, can function quite as well as it does, without “reference” in the philosopher’s sense, i.e. without referring to “objective” entities or events outside the experiential worlds of the members of a linguistic community. Hence, the fact that my neighbor “understands” what I want when I ask him for the carafe of water merely requires that the subjective experience my neighbor calls “carafe” is sufficiently like the subjective experience I myself have associated with that word. This does not entail that ontological reality contains a replica of either his or my carafe experience it merely presupposes a “reality” that is rich and ample enough so that both he and I can construct such experiences in the given context.
In other words, there is no reason to assume that his carafe experience and mine are “the same” all that is needed for my request to be understood, is that what he constructs as “carafe” satisfies the constraints I myself had in mind when I used the word. And if no match but only fit is required between his experience and mine, it would be downright absurd to assume that there should be a match between both our experiences, on the one hand, and an object in ontological reality, on the other.
Others and the notion of objectivity
This second illusory assumption springs from a habit of thought that is so deeply ingrained that we tend to take it as an “obvious” common sense fact: we consider our construct of Others to be unquestionable and ontologically real. Yet, how could it be? Our knowledge of Others, like all “the furniture of our experiential world, must originally have been composed out of elements of our own experience. I have elsewhere tried to give an account of how a cognizing organism may come to attribute the capabilities of perception, representation, cognition, and goal-directed action to certain items in the experiential field (von Glasersfeld, 1981, 1982). Once they have been made, these attributions may seem to be based on (and therefore confirm) a mystical belief or metaphysical conviction. In a rational model of the cognizing subject, however, neither of these sources would be admissible, because, by definition, they preclude any further rational analysis. If we want to model the knowing subject, we should avoid all presupposition of ready-made, intrinsic knowledge.
The notion of Others in the sense of particular items to whom one concedes capabilities similar to those that one attributes to oneself this notion can be grounded in the relatively simple realization that it may be advantageous to create a class of rather special items in one’s experiential field, namely items whose actions seem easier to predict and to control if one does assume that they have some of these prized capabilities. A butterfly perched on a flower in the field, for instance, will be easier to catch if one hypothesizes that it, too, can see and, therefore, will react to quick movements or changes of light and shadow in its immediate environment. Similarly, if I speak in order to induce my neighbor to pass the carafe of water, I must hypothesize that the speech sounds I emit will be perceived by him and will trigger certain cognitive operations and, eventually, motor acts that are somewhat compatible with those someone else’s utterance of these speech sounds would trigger in me. Hypotheses of that kind have worked fairly well in my past experience, and thus I have come to construct Others more or less in the image of myself.
This is by no means a new idea. Kant, in the first edition of his Critique of Pure Reason (1781), has a passage that expresses it with perfect clarity (p.223):
It is manifest that, if one wants to imagine a thinking being, one would have to put oneself in its place and to impute one’s own subject to the object that one intended to consider ...
(Remark: I am indebted to Eduard Marbach for having made me aware of this seminal passage in Kant’s early work. The translation is mine.)
What I find particularly seductive about this idea of imputing the model of oneself to Others, is that it provides a starting-point for an analysis of the concept of “objectivity”. From the constructivist perspective, as I explained earlier, “knowledge” comprises those constructs which the acting and knowing subject finds useful or at least tenable in the face of further experience. Clearly, then, if some of these constructs turn our to be viable not only in one’s own organization of experience but also as hypothetical basis for the computations and calculations one imputes to Others whose actions one wants to explain, predict, and control well, then one will almost inevitably come to think that these constructs are less “subjective” than those which one can find instantiated only in one’s own operating. As a constructivist, then, I can agree that knowledge should be called “objective” if it serves not only me, the subject, but also my interpretations of Others and their actions and understandings. But this does not and cannot warrant the assumption that, because a cognitive structure turns out to be useful in the interpretation of Others, it must on that account reflect, depict, or convey anything about the structure of an ontological reality.
A constructivist approach to ethics
The introduction of the concept of Others and, with it, of the level of experiential reality on which “objectivity” becomes operative, must sooner or later raise the question of ethics. At first it might, indeed, seem that no ethics whatsoever could be founded on an epistemology that is so explicitly centered on the cognizing subject. But this, I believe is not necessarily the case. Let me say at once that it is only recently that I have begun to think about this problem and what I have to offer is at best the suggestion of a starting-point.
The main difficulty in attempts to provide a rational grounding for ethics has always been that, whatever rules of conduct one wanted to justify, their justification required the assumption and acceptance of certain values and the scale on which these values were to be assessed invariably turned out to be itself in need of justification. In the con-structivist epistemology, I believe, it can be shown that at least a rudimentary ethics can be logically developed out of the very same assumptions that underly the central notion of the construction of knowledge.
This development can be summarized in the following steps:
(1) The model of the cognizing organism involves the working hypothesis that the organism’s intelligence is essentially self-regulating and aims at coordinating its experiences in such a way as to render experience explicable and manageable.
(2) Insofar as these coordinations or constructions are successful or viable, they constitute the organism’s experiential reality.
(3) Since criteria of viability vary with the context, there are levels of reality (repetition, mutual corroboration of sensory modes, confirmation by Others).
(4) The constructs that constitute this subjective reality will be considered “objective” if they turn out to be viable also in the construction of the cognizing subject’s models of Others (i.e. other cognizing subjects).
In order to achieve the highest level of “reality”, therefore, the cognizing subject not only needs Others but must also construe these Others with concepts that are not incompatible with those used in the construction of him- or herself; and, in order for these concepts to be and to remain viable not only for oneself but also for Others, one must necessarily assume that these Others operate within a goal structure that could conceivably be one’s own. It thus becomes clear that what Kant proposed as his Categorical Imperative is not merely an ethical prescription but is, i.n fact, a requirement of the individual’s own construction of a viable “objective” reality.
Both formulations of the Categorical Imperative are equally pertinent (Kant, 1788). “Act always in such a way that the guideline of your action could be taken as guideline by all Others”, means precisely that the models you construct of Others, in order to serve you as corroboration of your own reality, must be potential analogs of yourself, at least with regard to their goals and to the operations used to attain them. Thus, if the model you construct of yourself is to remain a viable one, the goals and operations you choose to construct it must always remain compatible with those which turn out to be viable in the construction of your models of Others.
As for Kant’s other formulation, “Treat Others as ends in themselves rather than as means to your own ends,” yields the interpretation that, in order to establish the highest level of your own reality, you must concede to Others the need and the possibility to construct their own reality.
The tentative claim I should like to make is clearly not that here is the possibility of an ethics that can do without assuming any a priori value. Rather, I would make the much more modest claim that constructivism may point the way to developing an ethics that requires no further assumptions than those that are inherent in the constructivist theory of knowledge.
It has taken many years to clarify the psychological feasibility of the processes of conceptual construction that are indispensable for a constructivist theory of knowledge. That such a clarification was possible at all is due to the pioneering work of Jean Piaget, who was the first to attempt an operational analysis of the mind. Given the revolutionary direction of the step he took in positing the human mind as a mechanism of self-regulation, it is of relatively little importance whether or not the actual operational analyses he produced can, in every detail, attain “objectivity” in the sense I have suggested above. Even if his contribution should eventually be reduced to the mere launching of a cognitive psychology that treats the organism as an informationally closed system that works to maintain its equilibrium in the flow of experiential perturbations, it would still have the enormous merit of having introduced a new and promising basis for the development of a non-contradictory theory of knowledge.
In order to survive, the constructivist theory of knowledge must be able to withstand attack. The present task, therefore, is to devise a reasonable defense against the philosophers’ almost immediate objection that any such constructivism is merely a new mask for the spectre of solipsism. This defense hinges on the change in the conception of “knowledge”. The notion of cognitive constructs that remain viable in the face of further experience, leads to a conception of “knowledge” which, though subject-generated, cannot be brushed aside as idealism because it does take into account the existence of an ontological reality. This ontological reality, however, is no longer a reality to be known, but rather a reality that constrains the range and the success of all acting and cognitive constructing.
This peculiar negative relation of knowledge to reality should not come as a great surprise to cyberneticians. Any self-regulating device, after all, “knows” only what it senses, and “acts” only when what it senses does not fit or satisfy the conditions or patterns that have been chosen as reference. In any cybernetic gadget, therefore, equilibrium is achieved whenever an interpretation of the sensory signals conforms to a pre-established desired pattern. Similarly, in the cognitive subject, equilibrium is achieved whenever the experiential situation can be satisfactorily managed without reorganization of the relevant conceptual structures.
The constructivist theory of knowledge is explicitly and unashamedly “instrumentalist”. Yet it should be immune to the usual arguments that attack instrumentalism because of its traditional connection with utilitarianism. For the constructivist, knowledge is not an instrument in the struggle for material benefits in an independent “objective” reality, but an instrument of equilibration in the cognizing subject’s experiential world.
Kant, I., 1781, Kritik der reinen Vernunft, 1. Auflage, (Gesammelte Schriften, Bd.IV). Koenigl. Preussische Akademie, 1910ff., Berlin.
Kant, I., 1788, Kritik der praktischen Vernunft, (Gesammelte Schriften, Bd.V). Koenigl. Preussische Akademie, 1910ff., Berlin.
McCulloch, W.S., 1965, Through the den of the metaphysician, in: “Embodiments of Mind”, MIT Press, Cambridge, Mass.
Pask, G., 1969, The meaning of cybernetics in the behavioral sciences, in: “Progress of Cybernetics”, J.Rose, ed., Gordon and Breach, New York.
Piaget, J., 1937, “La construction du réel chez l’enfant”, Delachaux et Niestle, Neuchatel.
Power, W.T., 1978, Quantitative analysis of purposive systems: Some spadework foundations of scientific psychology, Psychol.Rev., 85:417.
von Glasersfeld, E., 1981, Einführung in den radikalen Konstruktivismus, in: “Die erfundene Wirklichkeit”, P. Watzlawick, ed., Piper, Munich.
von Glasersfeld, E., 1982, An interpretation of Piaget’s constructivism. Revue Internationale de Philosophie, 36:612.
von Glasersfeld, E., 1984 Thoughts about space, time, and the concept of identity, in: “Book-Conference. First Instalment”, A. Pedretti, ed., Princelet Editions, London/Zurich.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/1389 on 2016-05-21 · Publication curated by Alexander Riegler